Research On Fast 3D Measurement Technology Combining Structured Light And Passive Light

Three-dimensional measurement technology is designed to obtain the 3D information.This technology has been widely applied in many fields.How to make a quick and high precision measurement has been a research focus for three dimensional measurements.Optical based 3D measurement technology is one of the most promising research directions.In general this technology can be divided into two aspects:passive and active.Passive stereo vision technology which usually only needs two cameras have advantages of simple measuring equipment,while the measuring accuracy is not high and it is always affected by illumination of environmental light.Active structured light measurement method has high accuracy,and in theory it is not affected by ambient light interference.While this technology needs additional coded patterns projected by projecting system,thus making projector calibration and phase wrapping become necessary.So to mitigate problems associated with passive stereo or actively structured light methods,the natural approach is to combine these two methods together.This thesis presents a new fast binocular stereo algorithm combining passive stereo matching and active structure light technology.First,passive stereo algorithm is used to get the initial disparity map.Then phase information of the active structured light is used to refine the initial disparity.And last,subpixel matching results are obtained.The main research contents and results are as follows:(1)Passive stereo vision and active structure light measurement techniques are studied respectively.After comparing several passive stereo algorithms,ELAS algorithm is used to get initial disparity.In active structure light,in order to achieve rapid measurement Fourier transform is used to obtain the phase information.(2)A new fast binocular stereo algorithm combining passive and active 3D reconstruction technology is put forward.First of all,a local phase matching algorithm is put forward to realize phase match pixel by pixel.Then by studying phase characteristics,subpixel level density matching results are realized through linear proportional method.Finally,in order to enhance the algorithm robustness,this paper proposes a window mean matching algorithm.(3)The advantages of proposed algorithm are as follows:Algorithm does not require any phase unwrapping process and is able to get sub-pixel accurate matching results.Besides the algorithm is strong robust.During the process of measuring,only one coded pattern is required achieving rapid measurement.No projector calibration and high quality phase map are needed.Finally,the proposed method can be applied to continuous or discontinuous objects.